We previously reported that human platelets aggregate spontaneously when decompressed to 1/3 atmosphere (Thrombosis Res. 33: 477, 1984) and that fibrinogen-dependent DIPA (decompression-inducible platelet aggregation) is inhibited by 10mM menthol, 10mM menthone, or 5 muM thymol (Thrombosis Res. 42:511, 1986). It is suggested that oppositely charged amino acid residues of fibrinogen are juxtaposed in the specific binding site(s). Under compression, electrically charged groups gain layers of water molecules which are pulled in closer to the center of the charge, making the layer more compact. This increases the thickness of the shield, disaggregating the platelets. Conversely, decompression allows the water to diffuse away from the compact shield; water molecules are randomized into the bulk phase; the shield becomes thinner, permitting spontaneous aggregation of platelets. Increasing the concentration of neutral salts inhibits platelets aggregation because high ionic strength attenuates the electric charge and thickens the shield. Molecules having a large electric dipole moment inhibit platelet aggregation by a similar mechanism. The theory has been experimentally confirmed (Thrombosis Res. 45; 729, 1987). Electric constriction appears to be involved in fibrin(ogen) dependent DIPA. The tetrapeptide R.G.D.S. (Arg- Gly-Asp-Ser) is found in both the specific binding site on the membrane surface of platelets as well as on the alpha-chain of fibrinogen. Synthetic tetrapeptide R.G.D.S. inhibits DIPA; 3 muM solution R.G.D.S changes its electric conductivity as a function of pressure, indicating dissociation under compression and association under decompression. The juxtaposed amino acid residues may be positively charged Arg and negatively charged Asp. Two alpha-helices were constructed from atom models, showing the non-convalent interactions between juxtaposed and oppositely charged amino acid residues consist of hydrogen- bonding with electrostatic interaction superimposed. This finding differs from the commonly held notion that fibrin polymerization takes place by way of H-bonding the hydrophobic interaction. There is no experimental indication of any hydrophobic interactions in DIPA and fibrin polymerization.